CN116610311A - Method and system for automatically excavating and multiplexing low-code templates - Google Patents
Method and system for automatically excavating and multiplexing low-code templates Download PDFInfo
- Publication number
- CN116610311A CN116610311A CN202310899398.6A CN202310899398A CN116610311A CN 116610311 A CN116610311 A CN 116610311A CN 202310899398 A CN202310899398 A CN 202310899398A CN 116610311 A CN116610311 A CN 116610311A
- Authority
- CN
- China
- Prior art keywords
- template
- code
- recommendation
- low
- user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000005065 mining Methods 0.000 claims abstract description 22
- 238000002372 labelling Methods 0.000 claims description 45
- 238000012549 training Methods 0.000 claims description 39
- 230000006399 behavior Effects 0.000 claims description 26
- 230000003993 interaction Effects 0.000 claims description 22
- 239000012634 fragment Substances 0.000 claims description 8
- 238000012986 modification Methods 0.000 claims description 7
- 230000004048 modification Effects 0.000 claims description 7
- 238000012937 correction Methods 0.000 claims description 6
- 230000014509 gene expression Effects 0.000 claims description 6
- 238000011161 development Methods 0.000 abstract description 9
- 230000006870 function Effects 0.000 abstract description 9
- 230000000694 effects Effects 0.000 abstract description 6
- 238000005516 engineering process Methods 0.000 description 7
- 230000007246 mechanism Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000013473 artificial intelligence Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/30—Creation or generation of source code
- G06F8/36—Software reuse
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
- G06N3/0455—Auto-encoder networks; Encoder-decoder networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Databases & Information Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a method and a system for automatically mining and multiplexing a low-code template, relates to the technical field of computer software information, and solves the problems that an existing low-code template cannot realize automation, large-scale and accurate understanding of scenes and functions of the template and actual application scenes of the template cannot be judged in application, and finally a recommendation effect is poor. The repeated labor is reduced, and the development efficiency is improved.
Description
Technical Field
The invention relates to the technical field of computer software information, in particular to a method, a system, equipment and a storage medium for automatically mining and multiplexing a low-code template.
Background
At present, the low code template automatic mining and multiplexing technology can improve the low code development efficiency, so that the method is widely applied. However, in the practical application process, the following modes are adopted to realize the process, and meanwhile, a plurality of problems are brought:
1. the template marking depends on a manual code template, and the marking process depends on manual classification and label input, so that a certain amount of labor is needed, and the accuracy of a marking result directly influences the follow-up recommendation effect due to the manual marking mode. If the labeling is not accurate enough, the accuracy of the recommendation is affected.
2. The template recommendation is difficult to be highly intelligent, and although a recommendation algorithm can be continuously optimized by using feedback by a developer, the template recommendation reaches an extremely high artificial intelligence level, the personalized and accurate recommendation of the developer is realized, the difficulty is relatively high, and the recommendation result in part of scenes does not necessarily completely meet the requirements of the developer.
3. Code templates with different compatibility problems may use different technical stacks and frameworks, which may cause that templates recommended by a recommendation system cannot be completely compatible, and a developer is required to perform additional work to make the templates work cooperatively, so that development difficulty is increased.
4. After the template maintenance cost code template library is continuously increased, a certain cost is required to be input for managing, maintaining and upgrading the template, and the template is required to be updated by continuously following the change of each technical stack and the frame by the platform, so that a certain burden is also increased for the platform.
5. Template reusability is difficult to guarantee, namely: not all code segments or items have strong reusability, and some templates have narrow applicable scenes. If recommended to an inapplicable scenario, the developer needs additional work modification and reconstruction, which increases the workload and affects the development efficiency.
In summary, for the implementation manner of the prior art, the management and mining of the code templates are mainly implemented by manual labeling and classification, mining of templates based on source code analysis, mining of high-quality templates based on developer behaviors, and the like, wherein a developer manually labels and classifies the code templates in detail, and then searches and filters the code templates with the help of labeling information. This requires a lot of labor cost and time, and automation and large scale cannot be achieved.
Mining templates based on source code analysis: by analyzing the similarity and structure between the source codes, the code templates are automatically mined. However, the actual scene and function of the template cannot be accurately understood, so that the recommendation effect is poor.
Excavating a high-quality template based on developer behaviors: the influence and the liveness of the code template are judged by analyzing the using and downloading behaviors of a developer, but the actual application scene of the template cannot be judged, and the result is recommended on one side.
Therefore, there is a need to develop a method and system for automatic mining and multiplexing of low-code templates.
Disclosure of Invention
The invention provides a method and a system for automatically excavating and multiplexing a low-code template, which are used for solving the problems that the conventional low-code template cannot realize automation and large-scale, cannot accurately understand the scene and function of the template, cannot judge the actual applicable scene of the template, and finally causes poor recommending effect.
A method for automatically mining and multiplexing a low-code template is realized by the following steps:
step one, acquiring code fragments to form a code file, and constructing a code template library of a low code platform by the code file and a code template generated by a user of the low code platform;
setting labeling rules, extracting notes and metadata information in codes of the code template library by adopting regular expressions as labels, and generating classification training data through manual labeling and content correction;
thirdly, performing multi-label classification training of template content by adopting a network based on ALBERT and textCNN according to the classification training data obtained in the second step, and generating a template classification labeling model after iteration for a plurality of times;
step four, the template classification marking model generated in the step three is adopted to carry out scale template marking, so that automatic classification marking of the code template is realized;
updating the number of templates and the template labels in the code template library, and recording interaction information between a user and the code templates;
step six, generating a training data set of the recommendation system by the interaction information recorded in the step five, the label information of the template mark and the user information;
performing code template recommendation training by adopting a recommendation model based on a transducer, and generating the code template recommendation model after iterating for a plurality of times;
and step seven, the user inputs search information by adopting the code template recommendation model generated in the step six, a user recommendation list is generated by using the code template recommendation model, and the user recommendation list is updated by using a listwise-based ranking algorithm.
The invention also provides a system for automatically excavating and multiplexing the low-code template, which comprises a code grabbing module, a template marking module, a template training and updating module, a behavior interaction module, a recommendation system generating module and a recommendation list updating module;
the code grabbing module is used for acquiring code fragments and forming a code file; constructing a code template library of a low code platform by the code file and a code template generated by a low code platform user;
the template marking module is used for: by adopting an artificial template labeling mode, extracting annotation and metadata information in codes as labels by setting specific labeling rules and assisting regular expressions, and finally generating classification training data by artificial labeling and content correction;
the template training and updating module: performing multi-label classification training of template content by using ALBERT and textCNN network methods, and obtaining a template classification labeling model after iteration for a plurality of times; performing large-scale template marking through the generated template classification marking model; updating the number of templates in the template library and the template labels;
the behavior interaction module is used for: recording interaction behaviors between a user and a code template, including, but not limited to, code template modification, code template browsing, code template use, code template sorting and the like, generating a data file and storing the data file;
generating a training data set of a recommendation system through interaction between a user and template behaviors, and label information and user information marked by the template;
the recommendation system generation module: performing template recommendation training by using a recommendation model based on a transducer, and generating a template recommendation model after iterating for a plurality of times;
the recommendation list updating module: and the user inputs search information into the template recommendation model, generates a recommendation list, and updates the recommendation list by adopting a ranking algorithm based on a list method.
The invention also provides an electronic device, which comprises a memory and a processor;
the memory is used for storing executable instructions;
and the processor is used for realizing the method for automatically mining and multiplexing the low-code template when executing the executable instructions stored in the memory.
The invention also provides a storage medium which stores executable instructions for realizing the method for automatically mining and multiplexing the low-code template when being executed by a processor.
The invention has the beneficial effects that:
1. in the method, the automatic labeling is carried out by adopting the deep learning model instead of manual classification, so that the template labeling automation is realized, and the labor cost is greatly reduced.
2. The method comprises information such as functions, technologies, applicability and the like of templates, and realizes recommendation based on requirements, and is not limited to source codes or developer behavior analysis.
3. In the method, an automatic labeling technology is utilized to train a labeling model by adopting a deep learning algorithm, a bi-directional coding pre-training model based on an ALBERT (A Lite Bert: simplified coding and decoding model based on an attention mechanism) and a multi-label text classification technology of a text convolution network (TextCNN) are used for training a code template labeling model, and the generated classification labeling model has high classification accuracy and reasoning speed, so that large-scale template classification labeling is realized, massive high-quality labeling information is contained, and the method is also the basis for realizing high-quality recommendation.
4. In the method, the recommendation system constructed based on the transducer attention model is used, the behavior relation of the labeling information between different developers and the code templates can be dynamically learned, the generated recommendation template list is learned and optimized through a list method (Listwise Approach), and personalized code template recommendation can be rapidly and accurately performed on the developers according to scene description and requirements input by the developers.
5. In the method, a corresponding relation between a developer and a scene is established by introducing a transducer attention mechanism into a recommendation model, the use preference of different developers when templates are applied is mined, and personalized recommendation is carried out. The invention realizes large-scale code template marking by using an automation technology, and realizes high-quality template recommendation and reuse based on the large-scale code template marking. The method realizes deep understanding and description of templates through labeling information, and is far beyond source code or behavior analysis method. Personalized recommendation and sequencing also show the improvement of the recommendation effect.
6. In the method, the automatic labeling is realized by using an artificial intelligence technology, and the code template management and reuse modes which are more intelligent and efficient are realized based on the depth understanding and recommendation of the templates. Through establishing a code template library with rich labels, the association and recommendation between templates are realized, a developer is helped to quickly find an applicable code template, repeated labor is reduced, and development efficiency is improved.
Drawings
FIG. 1 is a flow chart of a method for automatic mining and multiplexing of low code templates according to the present invention.
FIG. 2 is a schematic block diagram of a system for automatic mining and multiplexing of low code templates according to the present invention.
Detailed Description
The first embodiment describes the present embodiment with reference to fig. 1, and a method for automatically mining and multiplexing a low-code template is implemented by the following steps:
1. code segment grabbing: capturing code fragments and templates through an API interface matched with a crawler or a Git, wherein captured data are stored in a local file;
capturing a code template through an API interface matched with a crawler or a Git, and storing captured data in a local file. In this embodiment, code segments and application templates written by different developers are collected, and a code template library is built. The method can be used for grabbing from an open source library such as Github and the like, and can also be used for extracting from codes written and released by a developer on a low-code platform.
2. Code template library: the obtained code file and the code template generated by the user of the low code platform form a basic code template library of the low code platform together.
3. Marking an artificial template: annotation and metadata information in the regular expression extraction codes are used as labels by setting specific labeling rules, and finally classification training data are generated by manual labeling and content correction.
4. Template annotation model training: and performing multi-label classification training of template contents by using a network method based on ALBERT and textCNN, and obtaining a stable template classification labeling model with high accuracy after multiple iterations.
5. Automatic classification and labeling of code templates: performing large-scale template marking through the generated template classification marking model (namely, generating classification training data through manual template marking and training through the template classification marking model) to realize automatic classification marking of the code template;
labeling templates in a code template library, comprising: the method comprises the following steps: such as mobile terminal page development, web management background development, desktop application development, etc.; realizing functions such as data display, form submission, business flow and the like; technical stacks such as React, vue, java; other metadata, developer, date, score, etc.
6. Updating a code template library: continuously updating the number of templates in a template library and label information of template labels;
7. user template behavior interaction: recording interaction behaviors between a user and a code template, including, but not limited to, code template modification, code template browsing, code template use, code template sorting and the like, generating a data file and storing the data file;
8. recommendation system training data set: generating a training data set of a recommendation system through user template behavior interaction, and label information and user information of template labels;
9. training recommendation system: template recommendation training is carried out by using a recommendation model based on a transducer, and a code template recommendation model with stability, high accuracy and high coverage rate is generated after iteration is carried out for a plurality of times.
In this embodiment, the recommendation model based on the transducer is a machine learning model, which is used for personalized recommendation in the recommendation system. The model is built based on the attention mechanisms and can learn the historical behavior and preferences of the user to predict items or content that may be of interest to the user and recommend related items or content to the user. In code template recommendation, the model may learn the user's code needs and preferences through an attention mechanism, thereby predicting the code templates that the user may need and recommending relevant code templates to the user. Thus, the code template recommendation model represents a machine learning model that automatically recommends relevant code templates based on the needs and preferences of the user. 10. User retrieval generates a recommendation list: and the user inputs the search information, generates a recommendation list through the model, continuously updates the recommendation list by using a listwise-based ranking algorithm, and improves the recommendation accuracy.
In this embodiment, when a developer develops an application on a low-code platform, the platform may recommend a code template that matches information such as a code written by the developer, a selected technical stack, and a frame. The developer can select to use the recommended template, so that the repeated labor is reduced, and the development efficiency is improved.
After a developer uses a certain code template to modify or supplement the function of the code template, if the developer allows the code template, the low-code platform can extract the enhanced template of the developer and add the enhanced template into a code template library, so that the number of templates is continuously enriched, the quality of the templates is improved, and positive feedback is formed.
In this embodiment, the low-code platform may also open a code template library, allowing a developer to directly search and use a required template in the library, score the template, provide more accurate recommendation, and continuously optimize recommendation algorithms and mechanisms.
In the embodiment, the association between the code templates is established according to the scene and the function as the center, so that the scene and individuation of template recommendation are realized. The developer selects a certain scene or certain functions, and the low-code platform can recommend code template combinations containing relevant scenes and functions to quickly build an application prototype or framework.
The second embodiment is described with reference to fig. 2, and a system for automatically mining and multiplexing a low-code template is used for implementing the method described in the first embodiment, and the system includes a code grabbing module, a template marking module, a template training and updating module, a behavior interaction module, a recommendation system generating module and a recommendation list updating module;
the code grabbing module is used for acquiring code fragments and forming a code file; constructing a code template library of a low code platform by the code file and a code template generated by a low code platform user;
the template marking module is used for: by adopting an artificial template labeling mode, extracting annotation and metadata information in codes as labels by setting specific labeling rules and assisting regular expressions, and finally generating classification training data by artificial labeling and content correction;
the template training and updating module: performing multi-label classification training of template content by using a network method based on ALBERT and textCNN, and obtaining a template classification labeling model after iteration for a plurality of times; performing large-scale template marking through the generated template classification marking model; updating the number of templates in the template library and the template labels;
the behavior interaction module is used for: recording interaction behaviors between a user and a code template, including, but not limited to, code template modification, code template browsing, code template use, code template sorting and the like, generating a data file and storing the data file;
generating a training data set of a recommendation system through interaction between a user and template behaviors, and label information and user information marked by the template;
the recommendation system generation module: template recommendation training is carried out by using a recommendation model based on a transducer, and a code template recommendation model is generated after iteration is carried out for a plurality of times;
the recommendation list updating module: and the user inputs search information to the template recommendation model, generates a recommendation list, and updates the recommendation list by adopting a listwise-based ranking algorithm to improve the recommendation accuracy.
The system according to the present embodiment has the following advantages:
1. automatic labeling and recommendation of templates are realized by automatic and large-scale artificial intelligence technology, massive templates can be processed, the efficiency is greatly improved, and the limitation of manual classification and labeling is exceeded.
2. The labeling dimension is more comprehensive and accurate, and the code structure, annotation, dependence and the like are considered to generate accurate labeling content, so that manual omission or misjudgment is avoided.
3. Personalization and customization customizing template recommendation for developers according to skills and scenes of the developers exceeds a recommendation mechanism based on popularity and heat.
4. The template recommendation not only considers the similarity of source codes, but also recommends based on the actual use scene and purpose of the template, so that a more accurate recommendation effect is realized.
5. The continuous optimization algorithm can perform continuous optimization and iteration according to the evaluation feedback, and the recommendation quality is continuously improved. And the manual recommendation result is fixed once being produced, so that the manual recommendation result is difficult to optimize.
6. The standardization and consistency follow predefined standards and rules, so that consistency of labeling and recommendation among different templates is ensured to the greatest extent, and deviation introduced by manual habits is reduced.
7. The use threshold is reduced, professional manual classification and marking skills are not needed, marking and recommending rules are automatically generated by an algorithm, and the use threshold is remarkably reduced.
8. The continuous rich labeling content can not be updated in time with the increase of the number of templates, and the algorithm can analyze the newly added templates at any time to update the labeling content, so that the timeliness of the labeling content is ensured.
The third embodiment provides an apparatus, where the apparatus includes a memory, a processor, and a communication interface, where the memory, the processor, and the communication interface are electrically connected directly or indirectly to each other to implement data transmission or interaction. For example, the elements may be electrically connected to each other via one or more communication buses or signal lines.
The memory may be used to store software programs and modules, and the modules of the system described in the second embodiment may be stored in the memory in the form of software or firmware (firmware) or be solidified in an Operating System (OS) of the device, and the processor executes the software programs and modules stored in the memory, thereby performing various functional applications and data processing. The communication interface may be used for communication of signaling or data with other node devices.
Wherein the memory may be, but is not limited to, a random access memory (Random Access Memory,
RAM), read Only Memory (ROM), programmable Read Only Memory (Programmable Read-OnlyMemory, PROM), erasable Read Only Memory (Erasable Programmable Read-Only Memory, EPROM), electrically erasable Read Only Memory (Electric Erasable Programmable Read-Only Memory, EEPROM), and the like.
The processor may be an integrated circuit chip having signal processing capabilities. The processor may be a general purpose processor including a central processing unit (CentralProcessingUnit, CPU), a network processor (NetworkProcessor, NP), etc.; but also digital signal processors (Digital Signal Processing,15 DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
The fourth embodiment provides a readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method of the first embodiment. The computer readable storage medium may be, but is not limited to, a usb disk, a removable hard disk, ROM, RAM, PROM, EPROM, EEPROM, a magnetic disk, or an optical disk, etc. various media capable of storing program codes.
While certain exemplary embodiments of the present invention have been described above by way of illustration only, it will be apparent to those of ordinary skill in the art that modifications may be made to the described embodiments in various different ways without departing from the spirit and scope of the invention. Accordingly, the drawings and description are to be regarded as illustrative in nature and not as restrictive of the scope of the invention, which is defined by the appended claims.
Claims (8)
1. A method for automatically mining and multiplexing a low-code template is characterized by comprising the following steps: the method is realized by the following steps:
step one, acquiring code fragments to form a code file, and constructing a code template library of a low code platform by the code file and a code template generated by a user of the low code platform;
setting labeling rules, extracting notes and metadata information in codes of the code template library by adopting regular expressions as labels, and generating classification training data through manual labeling and content correction;
thirdly, performing multi-label classification training of template content by adopting a network based on ALBERT and textCNN according to the classification training data obtained in the second step, and generating a template classification labeling model after iteration for a plurality of times;
step four, the template classification marking model generated in the step three is adopted to carry out scale template marking, so that automatic classification marking of the code template is realized;
updating the number of templates in the code template library and label information of template labels, and recording interaction behaviors between a user and the code templates;
step six, generating a training data set of the recommendation system by the interaction behavior recorded in the step five, the label information of the template mark and the user information;
performing code template recommendation training by adopting a recommendation model based on a transducer, and generating the code template recommendation model after iterating for a plurality of times; finally, automatic mining and multiplexing are realized.
2. The method for automatically mining and multiplexing a low-code template according to claim 1, wherein: and step seven, the user inputs search information by adopting the code template recommendation model generated in the step six, generates a user recommendation list by using the code template recommendation model, and updates the user recommendation list by using a listwise-based ranking algorithm.
3. The method for automatically mining and multiplexing a low-code template according to claim 1, wherein: in the first step, capturing a code template through an API interface matched with a crawler or a Git, and storing captured data in a local file.
4. The method for automatically mining and multiplexing a low-code template according to claim 1, wherein: and fifthly, the interaction behavior between the user and the code template comprises code template modification, code template browsing, code template use and code template scoring, and a data file generated by the interaction behavior is stored.
5. A system for automatically mining and multiplexing a low-code template is characterized in that: the system comprises a code grabbing module, a template marking module, a template training and updating module, a behavior interaction module, a recommendation system generating module and a recommendation list updating module;
the code grabbing module is used for acquiring code fragments and forming a code file; constructing a code template library of a low code platform by the code file and a code template generated by a low code platform user;
the template marking module is used for: by adopting an artificial template labeling mode, extracting annotation and metadata information in codes as labels by setting specific labeling rules and assisting regular expressions, and finally generating classification training data by artificial labeling and content correction;
the template training and updating module: performing multi-label classification training of template content by using ALBERT and textCNN network methods, and obtaining a template classification labeling model after iteration for a plurality of times; performing large-scale template marking through the generated template classification marking model; updating the number of templates in the template library and the template labels;
the behavior interaction module is used for: recording interaction behaviors between a user and a code template, including code template modification, code template browsing, code template use and code template scoring, generating a data file and storing the data file;
generating a training data set of a recommendation system through interaction between a user and template behaviors, and label information and user information marked by the template;
the recommendation system generation module: performing template recommendation training by using a recommendation model based on a transducer, and generating a template recommendation model after iterating for a plurality of times;
the recommendation list updating module: and the user inputs search information into the template recommendation model, generates a recommendation list, and updates the recommendation list by adopting a ranking algorithm based on a list method.
6. The system for automatic mining and multiplexing of low-code templates according to claim 5, wherein: the code grabbing module collects code fragments written by different developers in the open source library or grabs the code fragments and the code templates from codes written and released by the developers on the low-code platform, and a code template library is built.
7. An electronic device, characterized by: comprising a memory and a processor;
the memory is used for storing executable instructions;
the processor, when configured to execute executable instructions stored in the memory, implements the method of any one of claims 1 to 4.
8. A storage medium, characterized by: executable instructions are stored for implementing the method of any one of claims 1 to 4 when executed by a processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310899398.6A CN116610311B (en) | 2023-07-21 | 2023-07-21 | Method and system for automatically excavating and multiplexing low-code templates |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310899398.6A CN116610311B (en) | 2023-07-21 | 2023-07-21 | Method and system for automatically excavating and multiplexing low-code templates |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116610311A true CN116610311A (en) | 2023-08-18 |
| CN116610311B CN116610311B (en) | 2023-12-05 |
Family
ID=87682262
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310899398.6A Active CN116610311B (en) | 2023-07-21 | 2023-07-21 | Method and system for automatically excavating and multiplexing low-code templates |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116610311B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190220253A1 (en) * | 2018-01-15 | 2019-07-18 | Cognizant Technology Solutions India Pvt. Ltd. | System and method for improving software code quality using artificial intelligence techniques |
| US20190391792A1 (en) * | 2018-06-25 | 2019-12-26 | Hcl Technologies Limited | Code reusability |
| CN114840252A (en) * | 2022-05-12 | 2022-08-02 | 中国工商银行股份有限公司 | Code annotation generation method, model training method, device and equipment |
| CN115145551A (en) * | 2022-08-02 | 2022-10-04 | 北京航空航天大学 | Intelligent auxiliary system for machine learning application low-code development |
| US20220334835A1 (en) * | 2021-08-16 | 2022-10-20 | Intel Corporation | Methods and apparatus to automatically evolve a code recommendation engine |
| CN115756475A (en) * | 2022-10-28 | 2023-03-07 | 浙江大学 | Sequence generation countermeasure network-based code annotation generation method and device |
| CN116028045A (en) * | 2022-12-08 | 2023-04-28 | 深圳复临科技有限公司 | Third party application code rapid distribution method, device, equipment and storage medium |
| CN116257644A (en) * | 2023-02-14 | 2023-06-13 | 阳光保险集团股份有限公司 | Method for marking data, method and device for marking data through model |
-
2023
- 2023-07-21 CN CN202310899398.6A patent/CN116610311B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190220253A1 (en) * | 2018-01-15 | 2019-07-18 | Cognizant Technology Solutions India Pvt. Ltd. | System and method for improving software code quality using artificial intelligence techniques |
| US20190391792A1 (en) * | 2018-06-25 | 2019-12-26 | Hcl Technologies Limited | Code reusability |
| US20220334835A1 (en) * | 2021-08-16 | 2022-10-20 | Intel Corporation | Methods and apparatus to automatically evolve a code recommendation engine |
| CN114840252A (en) * | 2022-05-12 | 2022-08-02 | 中国工商银行股份有限公司 | Code annotation generation method, model training method, device and equipment |
| CN115145551A (en) * | 2022-08-02 | 2022-10-04 | 北京航空航天大学 | Intelligent auxiliary system for machine learning application low-code development |
| CN115756475A (en) * | 2022-10-28 | 2023-03-07 | 浙江大学 | Sequence generation countermeasure network-based code annotation generation method and device |
| CN116028045A (en) * | 2022-12-08 | 2023-04-28 | 深圳复临科技有限公司 | Third party application code rapid distribution method, device, equipment and storage medium |
| CN116257644A (en) * | 2023-02-14 | 2023-06-13 | 阳光保险集团股份有限公司 | Method for marking data, method and device for marking data through model |
Non-Patent Citations (4)
| Title |
|---|
| TAM THE NGUYEN 等: "PERSONA：A personalized model for code recommendation", 《RESEARCH ARTICLE》, pages 1 - 15 * |
| WEIGUANG GONG 等: "Civil Aviation Travel Question and Answer Method Using Knoeledge Graphs and Deep Learning", 《ELECTRONICS2023》, vol. 12, no. 13 * |
| WENHUA WANG 等: "TranS3：A Transformer-based Framework for Unifying Code Summarization and Code Search", 《ARXIV:2003.03238V2 [CS.SE]》, pages 1 - 12 * |
| 林锴 等: "面向异常处理的代码智能化推荐", 《计算机科学与探索》, pages 667 - 677 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116610311B (en) | 2023-12-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109242013B (en) | Data labeling method and device, electronic equipment and storage medium | |
| CN111160569A (en) | Application development method and device based on machine learning model and electronic equipment | |
| US20200410392A1 (en) | Task-aware command recommendation and proactive help | |
| CN108845988B (en) | Entity identification method, device, equipment and computer readable storage medium | |
| CN111985229A (en) | Sequence labeling method and device and computer equipment | |
| CN112188311B (en) | Method and apparatus for determining video material of news | |
| US20220147702A1 (en) | Automatic generation of transformations of formatted templates using deep learning modeling | |
| CN111144079A (en) | Method and device for intelligently acquiring learning resources, printer and storage medium | |
| CN110929119A (en) | Data annotation method, device, equipment and computer storage medium | |
| US11501655B2 (en) | Automated skill tagging, knowledge graph, and customized assessment and exercise generation | |
| CN116594768A (en) | Large-model-oriented universal tool collaboration and refinement learning system and method | |
| CN116610304B (en) | Page code generation method, device, equipment and storage medium | |
| CN116610311B (en) | Method and system for automatically excavating and multiplexing low-code templates | |
| CN112836013A (en) | Data labeling method and device, readable storage medium and electronic equipment | |
| CN117236624A (en) | Issue repairer recommendation method and apparatus based on dynamic graph | |
| CN116822462A (en) | Method and device for generating simulation report of power system | |
| CN111488510A (en) | Method and device for determining related words of small program, processing equipment and search system | |
| CN116306506A (en) | Intelligent mail template method based on content identification | |
| Körner et al. | Mastering Azure Machine Learning: Perform large-scale end-to-end advanced machine learning in the cloud with Microsoft Azure Machine Learning | |
| CN115238653A (en) | Report generation method, device, equipment and medium | |
| CN111563371B (en) | Text generation method and device | |
| CN114625658A (en) | APP stability test method, device, equipment and computer readable storage medium | |
| CN114580533A (en) | Method, apparatus, device, medium, and program product for training feature extraction model | |
| CN113886580A (en) | Emotion scoring method and device and electronic equipment | |
| Zhang et al. | Big data-assisted urban governance: A comprehensive system for business documents classification of the government hotline |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |